3710

Hyperpolarized [1-13C] pyruvate as a biomarker for treatment monitoring in lymphoma murine models

Brice Tiret^1,2, Adam Autry², IlWoo Park², Min Lu³, Rana Anjum⁴, Lisa Drew⁴, Sebastien Degorce⁵, Michele Mayo⁴, Keith Dillman⁴, John Kurhanewicz², James Rubenstein³, and Myriam Chaumeil^1,2

¹Department of Physical Therapy and Rehabilitation Science, UCSF, San Francisco, CA, United States, ²Department of Radiology & Biomedical Imaging, UCSF, San Francisco, CA, United States, ³Helen Diller Family Comprehensive Cancer Center, UCSF, San Francisco, CA, United States, ⁴Oncology Innovative Medicines and Early Development, AstraZeneca, Waltham, MA, United States, ⁵Oncology Innovative Medicines, AstraZeneca, Cheshire, United Kingdom

Synopsis

In primary CNS lymphoma (PCNSL), MYD88 mutation is the most common, and induces a constitutive activation of IRAK4 and NF-kB signaling. AZ1495, a IRAK4 inhibitor, significantly improved OS in MYD88mut PCNSL models, highlighting the clinical potential of IRAK inhibition. However, monitoring of such targeted therapies remains challenging using conventional imaging. Because NF-kB controls metabolism, we questioned if hyperpolarized 13C MRSI could monitor IRAK inhibition in PCNSL. We show that HP-13C MRSI was able to detect a decreased HP lactate production following IRAK inhibition in MYD88mut tumors, not wild-types, highlighting the potential of metabolic imaging for monitoring therapy response in lymphomas.

Introduction

In primary central nervous system lymphoma (PCNSL), L265P mutation of the myeloid differentiation 88 (MYD88) protein is the most common¹. It induces a constitutive activation of IRAK4 kinase, IRAK1 phosphorylation and increased NF-κB signalling, a key player in cell survival, proliferation and metabolism, offering a unique window for therapeutic development. We previously showed that AZ1495, a potent inhibitor of IRAK4, significantly delayed progression and improved overall survival in L265P-mutated PCNSL models², highlighting the clinical potential of IRAK inhibition. In the clinic, however, early monitoring of such targeted therapies remains challenging using conventional imaging, as treatment often results in non-readily detectable metabolic changes. Because NF-κB controls metabolism as well, we questioned if hyperpolarized ¹³C magnetic resonance spectroscopic imaging (HP-¹³C MRSI) could non-invasively monitor IRAK inhibition in PCNSL harbouring MYD88 mutation.

Methods

Animals:

Two patient-derived models were studied: one mutant for MYD88 (MYD88mut), and one wild-type (MYD88wt). Mice (n=3 MYD88mut, n=5 MYD88wt) were injected intracranially with 200x10³ tumour cells on the left hemisphere. Xenograft bearing animals were imaged around 3 weeks post intracranial injection as the baseline (day 0, pre-treatment). At this point, animals received a daily dose AZ1495 treatment (50mg/kg/day per os) and were imaged at 2 and 4 post start of AZ1495 treatment.

MR acquisitions:

Each imaging session was performed on a 14.1T Varian system using a ¹H/¹³C dual tuned 40mm coil and Hypersense hyperpolarizer. T₂-weighted images were acquired using the following parameters: FOV 30mm, TR/TE 1200/20, 1 mm slice thickness. HP-¹³C MRSI was performed 17sec following injection of 24μL HP [1-¹³C] pyruvate using a FOV 32mm, TR 67ms, 4mm slice thickness, matrix 16x16.

Data analysis:

For T₂-w MRI, signal intensity was measured using ImageJ for the ipsilateral and contralateral sides. Ratios of the ipsilateral over the contralateral intensity was reported to account for possible changes in image gain. Manual segmentation of the whole brain was performed using 3DSlicer. For HP-¹³C MRSI data, the ratio of HP [1-¹³C] lactate to pyruvate was measured using SIVIC and MATLAB programs for the lesion site at day 0, 2 and 4.

Results

On T₂-w MRI (Fig.1A), no significant differences in signal intensity between groups at each time point were observed (p = 0.92, 0.34, 0.30 at days 0, 2 and 4 of treatment), nor between time points within each group (p= 0.79 and p=0.76 for MYD88mut and p= 0.28 and p=0.47 for MYD88wt), at day 2 and 4 of treatment respectively when compared with day 0 (Fig.1B). No changes in brain volumes were observed from baseline at days 2 and 4 (p=0.24 and p=0.32 for MYD88wt and p=0.74 and p=0.32 for MYD88mut respectively) (Fig.1C). In contrast, at day 0 of treatment i.e. 3 weeks post intracranial injection, high production of HP [1-¹³C] lactate was detected in both MYD88wt and MYD88mut (ratio±SEM 0.41±0.1 MYD88wt; 0.49±0.07 MYD88mut) (Fig.2). No significant changes in HP lactate-to-pyruvate ratios overtime were observed in the wild-type xenograft group at days 2 and 4 (p=0.53 and p=0.41 at days 2 and 4 respectively when compared to baseline). In contrast, in MYD88mut animals, the HP lactate-to-pyruvate ratios dropped significantly by 45% (p=0.02) at day 2 compared to baseline, but not at day 4 (p=0.4), in line with the previously reported effect of AZ1495 treatment in MYD88mut tumours only 2 (Fig.3). No significant changes between MYD88wt and MYD88mut groups were observed at days 0, 2 and 4 of treatment (p= 0.369, 0.151, 0.373 respectively).

Conclusion

To conclude, we showed that conventional T₂-w MRI cannot readily detect PCNSLs tumours, nor their response to targeted treatment; alternative imaging sequences need to be tested. On the other hand, HP-¹³C MRSI can detect MYD88-mutation-specific modulations of HP lactate production following IRAK inhibition. Because HP-¹³C MRSI is clinically translatable and expanding rapidly, this study is of high significance for future clinical trials, to enhance monitoring of PCNSL progression and response to targeted therapies.

Acknowledgements

No acknowledgement found.

References

1. Choi, J.W., et al., MYD88 expression and L265P mutation in diffuse large B-cell lymphoma. Hum Pathol, 2013. 44(7): p. 1375-81.

2. Geng, H., et al., Targeting NF-KB Activation in Novel Intracranial Models of CNS Lymphoma. Blood, 2016. 128(22): p. 777.

Figures

Figure 1. Conventional T₂-w MRI cannot detect PCNSLs nor their response to targeted therapy (A) Representative T₂-w MR image of the mouse brain and relative MRSI voxel size, 16x16 matrix in a 32mm FOV. (B) Signal intensity ratios between ipsilateral and contralateral side showing no changes in contrast between left (injection site) and right hemispheres in both groups. (C) Total brain volume measure for both the MYD88wt and MYD88mut group at day 0, 2 and 4 of treatment shows not significant differences between and within each group.

Figure 2. HP ¹³C MRSI can non-invasively detect PCNSLs xenograft and response to IRAK inhibition. Using HP-¹³C MRSI, a high HP lactate to pyruvate ratio due to the Warburg effect can be measured before the onset of treatment in booth MYD88mut and MYD88wt xenograft-bearing animals. In mice bearing the MYD88mut tumors, a significant decrease in lactate to pyruvate ratio can be observed at day 2 of treatment. No such decrease is observed in the mice bearing the wild type form of MYD88.

Figure 3. HP ¹³C-MRSI can non-invasively detect response to IRAK inhibition in MYD88-mutated PCNSLs xenogratfs. High production of HP [1-¹³C] lactate was detected in both MYD88wt and MYD88mut (ratio 0.41±0.1 MYD88wt; 0.49±0.07 MYD88mut). No significant changes in HP lactate-to-pyruvate ratios overtime was observed in the wild type xenograft group at days 2 and 4 (p=0.53 and p=0.41 at days 2 and 4 respectively). In contrast, in MYD88mut animals, the HP lactate-to-pyruvate ratios dropped significantly by 45% (p=0.02) at day 2 compared to baseline, but not at day 4 (p=0.4), in line with previously reported effect of AZ1495 treatment in MYD88mut tumors only.

Proc. Intl. Soc. Mag. Reson. Med. 26 (2018)

3710